PROJECT SUMMARY/ABSTRACT Genetically-encoded fluorescent biosensors allow us to capture real-time ?movies? of biochemical behavior inside individual cells, and they have already proven to be valuable tools for learning new biology. The most prominent examples are the intracellular calcium sensors, which reveal real-time neuronal activity in the intact brain, but there are many other new tools for measuring glucose concentration, protein kinase activity, caspase action, reactive oxygen species, and core metabolites such as ATP and NADH. Expressed via viral vectors or transgenes, they can be targeted to individual cells or cell types, and thus they can reveal time-dependent changes in signaling or metabolism in these cells in the context of a living, mixed-cell-type tissue ? and virtually all mammalian tissues are composed of multiple cell types with distinct roles in signaling and metabolism. In comparison, biochemical and mass-spec measurements have exquisite chemical sensitivity, but they usually involve sacrificing the preparation (making timecourses hard to learn), and like the also-powerful magnetic resonance spectroscopy/imaging technologies, they rarely have single-cell specificity. But unlike spectroscopic methods, the biosensors must be tailored specifically for each individual target. This generally involves a combination of semi-rational protein engineering ? in which a fluorescent protein and a ligand-binding protein are fused together in a specific way ? followed by screening of random or targeted mutagenic libraries of sensor variants. This screening process is a major limitation for sensor development, and a reason that many published biosensors are not adequately optimized ? meaning that many published sensors are a ?proof of principle? that cannot easily be used, or worse yet, have interferences that make them unreliable reporters of their nominal targets. One reason that optimization is challenging is that many characteristics of a sensor must be simultaneously optimized: the size of the fluorescence response, the sensitivity range for the target, the specificity of the sensor (including interference from other ligands), and resistance to environmental factors such as pH and temperature. We therefore aim to develop a high-throughput and high-content screening approach for genetically- encoded fluorescent biosensors, specifically for those that respond to ligand binding or other chemical stimuli. This screening method uses a series of well-established microfluidic and imaging methods, and we have piloted most of these already. When complete, this screening method should be deployable in other laboratories for widespread use. It will enable the screening of 104-105 sensor variants in less than a day, with information about each sensor variant in a dozen or more different conditions. We will also apply this screening approach to a series of published and unpublished biosensors in need of specific optimizations. This project will enable a dramatic improvement in the availability of high quality biosensors to study new biology.